24-hour electric power load prediction method

Abstract

The invention provides a 24-hour electric power load prediction method. The method comprises steps of data collection and data pre-processing, effective characteristics are calculated through the characteristic entropy weight, in combination with the window selection method, the 24-hour power load condition in the fixed time is selected as input data of DBN network training, the DBN network structure is determined, a network model is established, the network model is trained and tested through the DBN, characteristic value data of the prediction day and a 24-hour power load value of one day ofseveral days before the prediction day selected by the window selection method are inputted to obtain the electric power load value result of the prediction day, reasonable initialization of the weight is carried out through pre-training of the 2-layer RBM network layer, adjustment is carried out through the BP network layer, no shortcomings such as over-fitting occur, the characteristic entropyweight method is utilized to extract factors affecting the electric power load, the corresponding weight is calculated, and each factor is quantified to obtain the influence weight for the electric power load. The window selection method is proposed to improve prediction accuracy, and the prediction effect is better than an electric power load prediction model in the prior art.

Description

technical field [0001] The invention relates to the fields of power system planning and dispatching, and in particular to a method of using a DBN neural network model to perform 24-hour power load forecasting on a forecast day. Background technique [0002] With the continuous development of my country's electric power industry and the continuous improvement of people's living standards, the demand for electric energy in all walks of life is gradually increasing. Power load forecasting is the basis for formulating power generation plans and power system development plans. Accurate load forecasting is of great significance to the economical, safe and reliable operation of power systems. [0003] With the rapid development of modern science and technology, various load forecasting methods are constantly emerging. The load of the power system is uncontrollable and there is no comprehensive analysis considering that the influence factors such as temperature or weather have the ...

AI Technical Summary

Problems solved by technology

[0005] Aiming at the deficiencies in the prior art, the present invention provides a 24-hour power load forecasting method, the purpose of which is to solve the shortcomings of the BP neural network, such as the uncertainty in the selection of initial weights, overfitting, etc., and the prediction accuracy In terms of practicality and practicability, the requirements cannot be met at the same time. There is no comprehensive analysis and consideration of factors such as temperature or weather, which have periodic changes on a daily, weekly, and annual basis, which makes load forecasting inaccurate.

Images